metadata
dataset_info:
features:
- name: keys
dtype: string
- name: filename
dtype: string
- name: tags
sequence: string
- name: speaker_emb_path
dtype: string
- name: norm_lyrics
dtype: string
- name: recaption
dtype: string
splits:
- name: train
num_bytes: 3776000
num_examples: 32000
download_size: 68196
dataset_size: 3776000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*